INDEX

Explanations

phrases related to influential or top-ranking individuals

oai_token-act-pair · gpt-3.5-turbo

references to prominent individuals or groups in authority positions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ocene

-0.74

 Horses

-0.70

avia

-0.65

 Syndrome

-0.64

TPPStreamerBot

-0.64

cycles

-0.63

aris

-0.62

Interstitial

-0.62

crow

-0.62

Loading

-0.61

POSITIVE LOGITS

 diplomat

0.99

 aide

0.97

 prosecutor

0.96

 executive

0.94

 leadership

0.93

 negotiator

0.93

 member

0.92

 executives

0.92

 ministerial

0.92

 adviser

0.92

Activations Density 0.090%

phrases related to influential or top-ranking individuals

references to prominent individuals or groups in authority positions

No Comments

No Known Activations